Amazon Aurora is a fully managed relational database service designed for high performance, availability, and scalability. It is compatible with MySQL and PostgreSQL, offering the simplicity of open-source databases combined with enterprise-grade features and the reliability of AWS infrastructure.
High Performance:
- Up to 5x faster than standard MySQL and 3x faster than PostgreSQL.
- Uses a distributed, fault-tolerant, and self-healing storage system.
High Availability:
- Built-in fault tolerance with automatic recovery across multiple Availability Zones (Multi-AZ).
- Supports up to 15 low-latency read replicas within the same cluster.
Serverless Option:
- Aurora Serverless allows you to run databases without managing infrastructure.
- Scales automatically based on application demand.
Global Database:
- Enables fast, low-latency global read access by replicating data across multiple AWS Regions.
Automatic Backups:
- Continuous backups to Amazon S3 with point-in-time recovery.
Security:
- Integrates with AWS Identity and Access Management (IAM) for access control.
- Supports encryption at rest and in transit using AWS Key Management Service (KMS).
Durability
- Distributed, fault-tolerant, self healing storage system that scales up to 129TB per database.
Aurora Multi-Master
Aurora Multi-Master enables high availability and write scaling within a single AWS Region by allowing multiple database instances to handle read and write operations.
Multi-Writer Support:
All database instances in the cluster can process both read and write requests.
Useful for write-heavy applications.
Single Region Only:
Aurora Multi-Master works within a single AWS Region (cross-region writes are not supported).
Automatic Failover:
If one writer instance fails, another instance takes over without downtime or the need for promotion.
High Availability:
Eliminates the “single-writer bottleneck” by supporting multiple write nodes.
Consistency:
Leverages Aurora’s distributed storage system to ensure consistency across writer instances.
Use Cases:
- Write-heavy applications requiring write scalability and high availability.
- Applications needing minimal downtime for failover within a single region.
Limitations:
- Multi-Master is currently supported only for Aurora MySQL-Compatible Edition (not Aurora PostgreSQL).
- Can lead to potential contention or conflicts in applications with complex transactions or high concurrency.
- No cross-region write capability.
Aurora Global Database
Aurora Global Database is designed for applications requiring low-latency global reads, cross-region disaster recovery, and minimal downtime in case of regional outages.
Global Read Access:
Replicates data across up to 5 AWS Regions with latency of less than 1 second.
Provides fast, low-latency read operations for globally distributed applications.
Disaster Recovery:
The secondary region(s) can be promoted to primary within 1 minute in the event of a primary region failure.
Replication Type:
Uses asynchronous replication for cross-region replication to minimize latency.
Replication is highly efficient, only transferring database changes.
Independent Read Capacity:
Secondary regions can scale read replicas independently of the primary region.
Data Durability:
Leverages Aurora’s shared storage architecture, replicating data within each region across 3 Availability Zones.
Use Cases:
- Applications with a global user base requiring low-latency reads across multiple regions.
- Disaster recovery and business continuity for mission-critical workloads.
- Multi-region analytics, where secondary regions are used for analytics while the primary handles writes.
Limitations:
- Write operations are only allowed in the primary region (not multi-master).
- Applications needing multi-region writes must consider Aurora Multi-Master or other solutions.
